Report: US Watchful as EU Progresses With AI Policy
WASHINGTON — As the European Union moves ahead with its Artificial Intelligence Act, it’s time for American lawmakers to pay attention to potential policies, said John Soroushian, a senior associate director at the Bipartisan Policy Center.
“It’s important to pay attention to the interaction of AI and other technologies because AI doesn’t stand by itself,” Soroushian said in a phone interview Wednesday, days after he and his colleagues published a report on Europe’s pending policy.
“It’s important to look across the Atlantic as Europeans shape tech policy, specifically something like artificial intelligence that is increasingly used in everyday life,” Soroushian said.
“American lawmakers can take notes from their European counterparts, whom they often work closely with, and even if the U.S. doesn’t have the same laws, global tech products will possibly be made to fit European standards to appeal to the wider market,” he explained.
Artificial intelligence is a wide-ranging type of technology encompassing everything from voice-activated smartphones to devices that diagnose medical patients to banking software that decides on a loan.
It’s not easily defined and will evolve over time, which lawmakers will have to account for in potential laws, according to the report.
The technology, which in many cases is helpful, can also have negative consequences, according to the report.
“Harmful bias in AI systems can perpetuate or exacerbate historical inequities in areas such as employment, health care, finance, and housing, and malicious actors can use AI tools to manipulate people,” the report states.
And while many technology policy initiatives lag behind the creation of entire economic engines like social media, now — early in its adoption — is a great time for lawmakers to pay attention to how policy could help shape the technology, according to the report.
“AI is a fast-moving technology and public policy is something that’s usually slow to catch up,” Soroushian said. “But policy can help address many of the issues people have to help companies get ahead of them.”
So far, the European Union and the United States have differed in their approaches to regulation with Europe taking a more hands-on approach in this early stage.
“The European Commission has made it a priority to create a regulatory framework that would prevent and minimize AI’s negative effects. The United States, by contrast, has focused less on regulation and more on ‘soft law’ approaches, such as guidelines, standards, and frameworks, complemented by tort and other existing laws (such as civil rights laws) to help hold people liable when harm occurs,” according to the report.
The European policy model is also fairly subjective as a “risk-based” approach that sorts the level of regulatory scrutiny or all-out ban based on the risk.
“And there’s debate on what the risk is,” Soroushian said.
That’s why the proposed law is still pending after European Commission President Ursula von der Leyen announced she wanted to tackle AI regulations in 2019 after being elected. It was part of her first 100 days agenda to propose legislation on the issue.
At the end of her first 100 days, there was only a white paper exploring policy options. The draft legislation wouldn’t come out until 2021.
Europe is writing this AI law at the same time the union is proposing data privacy laws. The two intersect because data is key to AI technology, according to the report.
And American lawmakers are currently contemplating the first potential online data privacy law that would protect every resident; what is in that law will likely impact AI systems.
“Data is a key source for many modern AI systems, they are very reliant on data, so the kinds of data they can collect will affect how they are used. And it will create some kind of guardrails for that,” Soroushian said.
There are no specific recommendations in this report. Instead, the Bipartisan Policy Center is focusing on educating people on these issues, Soroushian said.
Taking note of what the EU and other countries are doing about AI will be important for American lawmakers to ascertain some of the best ways to implement legislation, he said. Other laws across the globe will likely impact the types of technology American consumers have access to, he added.
“The European Union has been doing a lot in this space, and the U.S. policymakers should be aware of this,” Soroushian said.